Using Hashtags to Capture Fine Emotion Categories from Tweets

Detecting emotions in microblogs and social media posts has applications for industry, health, and security. Statistical, supervised automatic methods for emotion detection rely on text that is labeled for emotions, but such data are rare and available for only a handful of basic emotions. In this article, we show that emotion‐word hashtags are good manual labels of emotions in tweets. We also propose a method to generate a large lexicon of word–emotion associations from this emotion‐labeled tweet corpus. This is the first lexicon with real‐valued word–emotion association scores. We begin with experiments for six basic emotions and show that the hashtag annotations are consistent and match with the annotations of trained judges. We also show how the extracted tweet corpus and word–emotion associations can be used to improve emotion classification accuracy in a different nontweet domain. Eminent psychologist Robert Plutchik had proposed that emotions have a relationship with personality traits. However, empirical experiments to establish this relationship have been stymied by the lack of comprehensive emotion resources. Because personality may be associated with any of the hundreds of emotions and because our hashtag approach scales easily to a large number of emotions, we extend our corpus by collecting tweets with hashtags pertaining to 585 fine emotions. Then, for the first time, we present experiments to show that fine emotion categories such as those of excitement, guilt, yearning, and admiration are useful in automatically detecting personality from text. Stream‐of‐consciousness essays and collections of Facebook posts marked with personality traits of the author are used as test sets.     

The Perils and Promises of Self-Disclosure on Social Media

In addition to their professional social media accounts, individuals are increasingly using their personal profiles and casual posts to communicate their identities to work colleagues. They do this in order to ‘stand out from the crowd’ and to signal attributes that are difficult to showcase explicitly in a work setting. Existing studies have tended to treat personal posts viewed in a professional context as a problem, since they can threaten impression management efforts. These accounts focus on the attempts of individuals to separate their life domains on social media. In contrast, we present the narratives of professional IT workers in India who intentionally disrupt the boundaries between personal and professional profiles in order to get noticed by their employers. Drawing on the dramaturgical vocabulary of Goffman (1959) we shed light on how individuals cope with increased levels of self-disclosure on social media. We argue that their self-presentations can be likened to post-modern performances in which the traditional boundaries between actor and audience are intentionally unsettled. These casual posts communicate additional personal traits that are not otherwise included in professional presentations. Since there are no strict boundaries between formal front-stage and relaxed back-stage regions in these types of performance, a liminal mental state is often used, which enables a better assessment of the type of information to present on social media.     

Predicting emotions in online social networks: challenges and opportunities

Online social networking has become a popular means of information exchange and social interactions. Users of these platforms generate massive amounts of data about their relationships, behaviors, interests, opinions, locations visited, items purchased, and subjective experiences of various aspects of life. Moreover, these platforms enable people from wide-ranging social and cultural backgrounds to synergize and interact. One interesting area of research is the emotional dimensions contained in this user-generated content, specifically, emotion detection and prediction, which involve the extraction and analysis of emotions in social network data. This study aimed to provide a comprehensive overview and better understanding of the current state of research regarding emotion detection in online social networks by performing a systematic literature review (SLR). SLRs help identify the gaps, challenges, and opportunities in a field of study through a careful examination of current research to understand the methods and results, ultimately highlighting methodological concerns that can be used to improve future work in the field. Hence, we collected and analyzed studies that focused on emotion in social network posts and discussed various topics published in digital databases between 2010 and December 2020. Over 239 articles were initially included in the collection, and after the selection process and application of our quality criteria, 104 articles were examined, and the results showed a robust extant body of literature on the text-based emotion analysis model, while the image-based requires more attention as well as the multiple modality emotion analysis.

     

Towards building a social emotion detection system for online news

Social emotion detection of online users has become an important task for mining public opinions. Social emotion detection aims at predicting the readers’ emotions evoked by news articles, tweets, etc. In this article, we focus on building a social emotion detection system for online news. The system is built based on the modules of document selection, Part-of-speech (POS) tagging, and social emotion lexicon generation. Empirical studies are extensively conducted on a large scale real-world collection of news articles. Experiments show that the document selection algorithm has a positive effect on the social emotion detection. The system performs better with the words and POS combination compared to a feature set consisting only of words. POS is also useful to detect emotion ambiguity of words and the context dependence of their sentiment orientations. Furthermore, the proposed method of generating the lexicon outperforms the baselines in terms of social emotion prediction.     

Personal summarization from profile networks

Personal profile information on social media like LinkedIn.com and Facebook.com is at the core of many interesting applications, such as talent recommendation and contextual advertising. However, personal profiles usually lack consistent organization confronted with the large amount of available information. Therefore, it is always a challenge for people to quickly find desired information from them. In this paper, we address the task of personal profile summarization by leveraging both textual information and social connection information in social networks from both unsupervised and supervised learning paradigms. Here, using social connection information is motivated by the intuition that people with similar academic, business or social background (e.g., comajor, co-university, and co-corporation) tend to have similar experiences and should have similar summaries. For unsupervised learning, we propose a collective ranking approach, called SocialRank, to combine textual information in an individual profile and social context information from relevant profiles in generating a personal profile summary. For supervised learning, we propose a collective factor graph model, called CoFG, to summarize personal profiles with local textual attribute functions and social connection factors. Extensive evaluation on a large dataset from LinkedIn.com demonstrates the usefulness of social connection information in personal profile summarization and the effectiveness of our proposed unsupervised and supervised learning approaches.     

Fuzzy model of dominance emotions in affective computing

To date, most of the human emotion recognition systems are intended to sense the emotions and their dominance individually. This paper discusses a fuzzy model for multilevel affective computing based on the dominance dimensional model of emotions. This model can detect any other possible emotions simultaneously at the time of recognition. One hundred and thirty volunteers from various countries with different cultural backgrounds were selected to record their emotional states. These volunteers have been selected from various races and different geographical locations. Twenty-seven different emotions with their strengths in a scale of 5 were questioned through a survey. Recorded emotions were analyzed with the other possible emotions and their levels of dominance to build the fuzzy model. Then this model was integrated into a fuzzy emotion recognition system using three input devices of mouse, keyboard and the touch screen display. Support vector machine classifier detected the other possible emotions of the users along with the directly sensed emotion. The binary system (non-fuzzy) sensed emotions with an incredible accuracy of 93 %. However, it only could sense limited emotions. By integrating this model, the system was able to detect more possible emotions at a time with slightly lower recognition accuracy of 86 %. The recorded false positive rates of this model for four emotions were measured at 16.7 %. The resulted accuracy and its false positive rate are among the top three accurate human emotion recognition (affective computing) systems.     

基于多传递影响力的社交媒体谣言检测方法

社交媒体谣言检测是当前研究的热点问题,现有方法多数通过获取大量用户属性学习用户特征,但不适用于谣言的早期检测,忽略了用户之间的潜在关系对信息传播的影响。提出一种基于多传递影响力的谣言检测方法,根据源微博及其对应转发（评论）之间的关系构建文本信息传播图,并通过图卷积神经网络来捕获、学习文本信息的传播特征。利用文本信息和用户传播过程中的影响力,丰富可用于谣言检测早期的检测信息。将存在转发关系的用户构成用户影响力传播图,构建一种用户节点影响力学习方法,获取用户节点影响力,以增强用户特征信息。在此基础上,将文本特征与用户特征融合以进行谣言检测,从而提升检测效果。在3个真实社交媒体数据集上的实验结果表明,该方法在谣言自动检测以及早期检测的效果都有显著提升,与目前最好的基准方法相比,在微博、Twitter15、Twitter16数据集上的正确率分别提高了2.8%、6.9%和3.4%。     

Your posts betray you: Detecting influencer-generated sponsored posts by finding the right clues

With the prevalence of sponsorship practice using social media posts, the detection of sponsored content becomes crucial for platforms to regulate the generated content and prevent users from being misinformed. However, there is a paucity of investigations on the detection of sponsored content in existing research. To fill this research gap, we first identify several task-related clues by referring to relevant psychological theories and practical observations. Based on the clues, we conceptualize four types of sponsored content features and propose a unified deep learning detection framework, which also learns the relative importance of each feature. Experiments conducted on 26,823 social media posts demonstrate the performance of our proposed model compared with competitive alternatives and the value of each feature. The learned feature importance also enables deeper phenomena understanding. The research findings provide actionable insights into the narrative strategies influencers adopt and how to distinguish online sponsored content.     

Film mood induction and emotion classification using physiological signals for health and wellness promotion in older adults living alone

This paper introduces a wearable hardware/software system specifically tailored to detect seven emotions (neutral, tenderness, amusement, anger, disgust, fear, and sadness) aimed at promoting health and wellness in older adults living alone at home. The complete software and hardware architectures acquiring and processing electrodermal activity and photoplethysmography signals are introduced. The wearable emotion detection system is trained by eliciting the desired emotions on 39 older adults through a film mood induction procedure. Seventeen features are calculated on skin conductance response and heart rate variability data, grouped into five statistical, four temporal, and eight morphological features. Then, these features are used to run emotion classification considering support vector machines, decision trees, and quadratic discriminant analysis. In line with psychological findings, the results offer a global accuracy of 82% in negative emotion (anger, disgust, fear, and sadness) classification. For positive emotions (tenderness and amusement), also in conformity with previous psychological outcomes, amusement shows the highest ratio of hits (92%) but tenderness the lowest one (66%). These results demonstrate that our wearable emotion detection system can be used by ageing adults, especially for detecting negative emotions that usually damage health and wellness and lead to social isolation.     

Suspended because of social media? Students’ knowledge and opinions of university social media policies and practices

In this exploratory study, we examined undergraduates’ (N = 298) knowledge of their university’s social media policies, understanding of free speech and privacy protections, opinions about university monitoring and discipline for personal social media posts, and perceptions of fairness regarding recent cases of student discipline for personal social media use. The results of our study indicate that most undergraduates are highly underinformed as to whether or not their university has a social media policy, particularly if the students are early in their academic careers and do not engage in many online privacy protection behaviors. Most participants were also misinformed as to whether free speech and/or privacy protections will shield them from university discipline. In addition, most participants (78%) were opposed to the idea of universities monitoring students’ personal social media accounts, though significantly fewer (68%) were opposed to monitoring student athletes’ social media. Finally, when asked about several recent cases involving student discipline, most participants were generally opposed to a variety of university disciplinary actions regarding students’ social media posts. We discuss these findings as they relate to the need for better social media policy training for students, as well as the potential impact on students’ academic and future careers.     

Social media crisis management: Aligning corporate response strategies with stakeholders’ emotions online

The article focuses on how the analysis of stakeholders’ emotions online can help companies facing a social media crisis determine the response strategy that will best minimize the reputational threat. The article indeed questions the relevance of classical crisis management theory to an online environment. Results show that social media have increased the unpredictability of corporate crises. Consequently, on social media, crises cannot be addressed with the methods that have prevailed so far. Rather, incorporating emotion‐based analysis in six case studies showed how crisis analysis, and the subsequent response strategy, could be fine‐tuned. The article builds on recent literature to develop a new analytical framework for response strategies and a model for crisis resolution—the social media crisis management matrix.     

Social Roles and Consequences in Using Social Media in Disasters: a Structurational Perspective

Disaster management officials, as well as the general public, are increasingly using social media to communicate. Such usage has resulted in new and emergent social consequences for disaster management and has reformed the roles of its relevant stakeholders. However, the existing literature on social media use in disasters is still preliminary and incomplete, and does not capture the change in social roles that stakeholders have taken and the consequences of the actions that people take in using social media. In this paper, by using Structuration theory as a meta-theory and by analysing the posts and comments in three officials’ Facebook fan pages in three different disasters, we theorize the social structures (i.e., social roles and social consequences) and the human actions taken by both the public and the disaster management officials during disasters. Furthermore, we explain how the social structures emerge out of the human actions involved, and how the social structures further shape those actions. Our research provides theoretical and practical insights into how the usage of social media in disasters benefits disaster management and reinforces the roles of the different stakeholders.     

Social reader: towards browsing the social web

We describe Social Reader, a feed-reader-plus-social-network aggregator that mines comments from social media in order to display a user’s relational neighborhood as a navigable social network. Social Reader’s network visualization enhances mutual awareness of blogger communities, facilitates their exploration and growth with a fully dragn- drop interface, and provides novel ways to filter and summarize people, groups, blogs and comments. We discuss the architecture behind the reader, highlight tasks it adds to the workflow of a typical reader, and assess their cost. We also explore the potential of mood-based features in social media applications. Mood is particularly relevant to social media, reflecting the personal nature of the medium. We explore two prototype mood-based features: colour coding the mood of recent posts according to a valence/arousal map, and a mood-based abstract of recent activity using image media. A six week study of the software involving 20 users confirmed the usefulness of the novel visual display, via a quantitative analysis of use logs, and an exit survey.     

Modeling user interest in social media using news media and wikipedia

Social media has become an important source of information and a medium for following and spreading trends, news, and ideas all over the world. Although determining the subjects of individual posts is important to extract users' interests from social media, this task is nontrivial because posts are highly contextualized and informal and have limited length. To address this problem, we propose a user modeling framework that maps the content of texts in social media to relevant categories in news media. In our framework, the semantic gaps between social media and news media are reduced by using Wikipedia as an external knowledge base. We map term-based features from a short text and a news category into Wikipedia-based features such as Wikipedia categories and article entities. A user's microposts are thus represented in a rich feature space of words. Experimental results show that our proposed method using Wikipedia-based features outperforms other existing methods of identifying users' interests from social media.     

基于异构社交上下文的多视图微博主题检测

社交媒体主题检测旨在从大规模短帖子中挖掘潜在的主题信息. 由于帖子形式简短、表达非正规化, 且社交媒体中用户交互复杂多样, 使得该任务具有一定的挑战性. 前人工作仅考虑了帖子的文本内容, 或者同时对同构情境下的社交上下文进行建模, 忽略了社交网络的异构性. 然而, 不同的用户交互方式, 如转发, 评论等, 可能意味着不同的行为模式和兴趣偏好, 其反映了对主题的不同的关注与理解; 此外, 不同用户对同一主题的发展和演化具有不同影响, 社区中处于引领地位的权威用户相对于普通用户对主题推断会产生更重要的作用. 因此, 提出一种新的多视图主题模型(multi-view topic model, MVTM), 通过编码微博会话网络中的异构社交上下文来推断更加完整、连贯的主题. 首先根据用户之间的交互关系构建一个属性多元异构会话网络, 并将其分解为具有不同交互语义的多个视图; 接着, 考虑不同交互方式与不同用户的重要性, 借助邻居级注意力和交互级注意力机制, 得到特定视图的嵌入表示; 最后, 设计一个多视图驱动的神经变分推理方法, 以捕捉不同视图之间的深层关联, 并自适应地平衡它们的一致性和独立性, 从而产生更连贯的主题. 在3个月新浪微博数据集上的实验结果证明所提方法的有效性.     

Dissecting emotion and user influence in social media communities: An interaction modeling approach

Human emotion expressed in social media plays an increasingly important role in shaping policies and decisions. However, the process by which emotion produces influence in online social media networks is relatively unknown. Previous works focus largely on sentiment classification and polarity identification but do not adequately consider the way emotion affects user influence. This research developed a novel framework, a theory-based model, and a proof-of-concept system for dissecting emotion and user influence in social media networks. The system models emotion-triggered influence and facilitates analysis of emotion-influence causality in the context of U.S. border security (using 5,327,813 tweets posted by 1,303,477 users). Motivated by a theory of emotion spread, the model was integrated in an influence-computation method, called the interaction modeling (IM) approach, which was compared with a benchmark using a user centrality (UC) approach based on social positions. IM was found to have identified influential users who are more broadly related to U.S. cultural issues. Influential users tended to express intense emotions of fear, anger, disgust, and sadness. The emotion trust distinguishes influential users from others, whereas anger and fear contributed significantly to causing user influence. The research contributes to incorporating human emotion into the data-information-knowledge-wisdom model of knowledge management and to providing new information systems artifacts and new causality findings for emotion-influence analysis.     

基于社交媒体的关联性用户属性推断

挖掘用户属性对用户建模、用户检索和个性化服务等具有十分重要的意义.已有的相关研究工作都是单独挖掘各种属性,而且忽略了各属性之间的相关关系.提出一种基于超图学习的用户属性推断的方法.在超图中,顶点表示社会媒体中的用户,超边表示用户产生的内容相似性与属性之间的关系.在建好的超图模型上,把用户属性挖掘形式化成一个正则化的标签相似传播问题,可以有效推断得到用户的各种属性.利用从Google+上收集的标记过全部属性的数据集进行了大量的实验,其结果表明了该方法在用户属性挖掘中的有效性.     

Tales of emotion and stock in China: volatility,causality and prediction

How the online social media, like Twitter or its variant Weibo, interacts with the stock market and whether it can be a convincing proxy to predict the stock market have been debated for years, especially for China. As the traditional theory in behavioral finance states, the individual emotions can influence decision-makings of investors, it is reasonable to further explore these controversial topics systematically from the perspective of online emotions, which are richly carried by massive tweets in social media. Through thorough studies on over 10 million stock-relevant tweets and 3 million investors from Weibo, it is revealed that inexperienced investors with high emotional volatility are more sensible to the market fluctuations than the experienced or institutional ones, and their dominant occupation also indicates that the Chinese market might be more emotional as compared to its western counterparts. Then both correlation analysis and causality test demonstrate that five attributes of the stock market in China can be competently predicted by various online emotions, like disgust, joy, sadness and fear. Specifically, the presented prediction model significantly outperforms the baseline model, including the one taking purely financial time series as input features, on predicting five attributes of the stock market under the K-means discretization. We also employ this prediction model in the scenario of realistic online application and its performance is further testified.     

Social media and relationship development: The effect of valence and intimacy of posts

Social media is becoming an increasingly common part of everyday life. Many social media sites (e.g. Facebook, Twitter and LinkedIn) support new interpersonal interaction methods, some of which are neither directed nor reciprocated. For example, social media users can read online 'posts' (self-disclosures) of their friends without interacting with those friends. This is vastly different to traditional face-to-face communication. Our study investigated how reading online 'posts' affects relationship development. Using a longitudinal design sampling 243 participants, we focused on the effect of the posts' valence and intimacy. We found that high intimacy posts or negative posts decreased the social attractiveness of the self-discloser. The perception of the posts and the receiver's feelings of homophily to the self-discloser mediated this relationship. Studies of offline interpersonal interaction have found similar results. In offline communication, self-disclosure perception and homophily also mediate relationship outcomes. This suggests that reading posts on social media and interacting in real life trigger similar or identical relationship formation pathways. These results support the argument that passive consumption is a new method of interaction that does not fundamentally change human psychology. While novel, passive consumption is still based on the same principles as offline communication.     

多模态多层次事件网络的谣言检测

目的 自动检测谣言至关重要,目前已有多种谣言检测方法,但存在以下两点局限：1）只考虑文本内容,忽略了可用于判断谣言的辅助多模态信息;2）只关注时间序列模型捕捉谣言事件的时间特征,没有很好地研究事件的局部信息和全局信息。为了克服这些局限性,有效利用多模态帖子信息并联合多种编码策略构建每个新闻事件的表示,本文提出一种新颖的基于多模态多层次事件网络的社交媒体谣言检测方法。方法 通过一个多模态的帖子嵌入层,同时利用文本内容和视觉内容;将多模态的帖子嵌入向量送入多层次事件编码网络,联合使用多种编码策略,以由粗到细的方式描述事件特征。结果 在Twitter和Pheme数据集上的大量实验表明,本文提出的多模态多层次事件网络模型比现有的SVM-TS（support vector machine—time structure）、CNN（convolutional neural network）、GRU（gated recurrent unit）、CallAtRumors和MKEMN（multimodal knowledge-aware event memory network）等方法在准确率上提升了4 %以上。结论 本文提出的谣言检测模型,对每个事件的全局、时间和局部信息进行建模,提升了谣言检测的性能。